The factors that regulate T cell memory are of great interest, and of potential significance to understanding how to augment immunity and how to suppress adverse immune reactions. Data gathered many years ago promoted the idea that the longevity of effective memory was dependent on periodic exposure to antigen {Gray, D., Matzinger, P. (1991) J Exp Med. 174:969). In contrast, more recent data have challenged this idea and suggested that individual memory T cells can survive for extended periods in the absence of specific antigen, and in some cases in the absence of any apparent signals (Lau et al., Nature (1994) 369:648; Tanchot et al., Science (1997) 276:2057; and Swain et al., Science (1999) 286:381).
Rather than one theory being incorrect, there are scenarios that may incorporate both ideas. For example, an alternative hypothesis is that a memory T cell can persist and be functional in the absence of signals accompanying antigen recognition, but that when antigen is encountered again, the individual memory T cell is now subject to both positive and negative signals that will dictate its further persistence and/or further functionality. It can be envisioned that some types of memory, such as that driving allergic asthma, are maintained in the face of antigen insults. Although it can be argued that other types of memory may not involve periodic exposure to antigen, at some stage those memory T cells will be required to respond to their specific antigen and hence again be susceptible to any positive and negative signals that accompany that recognition event. In either case, it is therefore essential to define the nature of those positive and negative signals.
Apart from the TCR/peptide/MHC interaction, the most likely sources of positive and negative signals are membrane bound molecules of the Ig and TNFR superfamilies. Although there is abundant data on the requirement of members of these families in effective priming of naïve T cells and hence dictating the development of memory T cells, there is virtually no data on whether they can influence reactivity or persistence of a memory T cell once it has been generated. For example, although it is widely accepted that the response of a naïve T cell is positively controlled by signals from cell surface costimulatory receptors, the role of costimulation in regulating a memory T cell has not been established. Based on experiments in vitro a number of years ago, it was postulated that memory T cells are less dependent, or independent, of costimulation for activation (Croft et al., J Immunol. (1994) 152:2675; Byrne et al., J Immunol. (1988) 141:3249; and Luqman, M., Bottomly, K. J Immunol. (1992) 149:2300). Negative data from in vivo studies trying to block CD28 costimulation from B7 also supported this idea (Lu et al., J Immunol. (1995) 54:1078; Gause et al., Exp Parasitol. (1996) 84:264; and Harris et al., Eur J Immunol. (1999) 29:311), while only a few publications have suggested that a secondary response may be susceptible to CD28 signals (Keane-Myers et al., J Immunol. (1997) 158:2042; and Tsuyuki et al., J Exp Med. (1997) 185:1671).
In contrast to costimulation through CD28, it is now clear that a number of additional receptors exist that may be crucial to a long-lived T cell response. The contribution of these other pathways to maintenance and functionality of antigen-specific memory T cells is unknown. OX40 (CD134) is one such costimulatory member, belonging to the TNFR superfamily. OX40 (CD134) has been shown to mediate potent costimulatory activity upon binding to its cognate ligand, OX40L, expressed on APC (Weinberg et al., Semin Immunol. (1998)10:471). OX40 is not constitutively expressed on naïve T cells but is induced 24-48 hr after recognition of antigen (Mallett et al., EMBO J. (1990) 9:1063; Calderhead et al., J Immunol. (1993) 151:5261; Baum et al., EMBO J. (1994) 13:3992; and Gramaglia et al., J Immunol. (1998) 161:6510. OX40L, a member of the TNF family, is also inducible being expressed on activated B cells, dendritic cells, and macrophage-like cells (Baum et al. (1994), supra; Gramaglia et al. (1998), supra; Al-Shamkhani et al., J Biol Chem. (1997) 272:5275; Stuber et al., Immunity (1995) 2:507; Ohshima et al., J Immunol. (1997) 159:3838; and Weinberg et al., J Immunol. (1999) 162:1818).
Previous work has demonstrated that OX40 and OX40L control the development of a number of primary T cell responses (Gramaglia et al. (1998), supra, Gramaglia et al., J Immunol. (2000) 165:3043, Rogers et al., Immunity (2001) 15:445, Weinberg et al., (1999), supra, Kopf et al., Immunity (1999) 11:699, Chen et al., Immunity (1999) 11:689, Murata et al., J Exp Med. (2000) 191:365, Jember et al., J Exp Med. (2001) 193:387, Akiba et al., J Exp Med. (2000) 191:375, Tsukada et al., Blood (2000) 95:2434, Higgins et al., J Immunol. (1999) 162:486, Yoshioka et al., Eur J Immunol. (2000) 30:2815, Nohara et al., J Immunol. (2001) 166:2108). OX40 appears to function by suppressing T cell death by maintaining high levels of anti-apoptotic proteins such as Bcl-xL and Bcl-2 (Rogers et al., Immunity (2001) 15:445) and inhibiting expression or activity of pro-apoptotic proteins such as Bad and Bim. This conclusion is supported by in vivo studies where agonist antibodies directed to OX40 on a responding naïve CD4 cell enhanced primary T cell expansion and survival, promoting the development of greater numbers of memory T cells (Gramaglia et al., (2000), supra; Maxwell et al., J Immunol. (2000) 164:107). However, no studies have addressed whether these interactions are required by memory T cells.
OX40 is down-regulated after the effector phase of primary T cell responses and returns to baseline levels within a week after initial antigen encounter. Significantly, antigen-primed CD4+ T cells can upregulate OX40 more rapidly than naïve T cells and the majority can re-express OX40 within four hours of antigen stimulation (Gramaglia et al. (1998), supra). Similarly, an anergic T cell, which also represents an antigen-experienced cell, albeit functionally hyporesponsive, can also re-express OX40 at low levels and be receptive to OX40 engagement resulting in enhanced functionality (Bansal-Pakala et al., Nat Med. (2001) 7:907).
Previous studies concluded that memory T cells largely had a reduced requirement for costimulatory signals and, therefore, may not be susceptible to interventions that target such membrane bound molecules as OX40 (Croft, M. Curr Opin Immunol. (1994) 6:431). For example, blocking B7-CD28 interactions during secondary responses to the nematode parasites Heligmosomoides polygyrus and Nippostrongylus brasiliensis, or immunogenic anti-mouse IgD antibody treatment, failed to inhibit memory Th2 responses, whereas blocking CD28 at the time of priming was suppressive (Lu et al., J Immunol. (1995) 54:1078, Gause et al., Exp Parasitol. (1996) 84:264, Harris et al., Eur J Immunol. (1999) 29:311). These and other observations therefore imply that activation of memory Th2 cells may be costimulation, or at least B7/CD28, independent.